How the monitored publications were assessed
Discourse valence across every scored item: what the sampled sources published and how independent models read it. It is not a measure of public opinion, not a measure of the state of the world, and article volume is not agreement.
This line is the mean of every item scored on each day. It deliberately ignores the reading filter, because a daily mean of only the strongly adverse items would move with how many were filtered out rather than with the discourse.
12 items read as strongly adverse
- “This is the AI men actually use”: Meta ads pushed apps nudifying real teens
- Anthropic researchers say AI could cause human extinction by 2030
- Anthropic details bad actors’ efforts to misuse its AI for bioweapons
- Top US official named to OpenAI non-profit board warns advanced AI could be ‘deadly’
- Lawmakers blast AI companies after researcher warns of human extinction by 2030
- We have started losing control of AI. It’s time to shut it down | Garrison Lovely
- Anthropic researcher quits over AI labs ‘gambling with our lives’
- AI could kill all humans in next decade, warn experts: but how seriously should we take them?
- Man told ChatGPT he was feeling delusional. ChatGPT insisted he was Jesus.
- OpenAI not on track to reduce risk of ‘catastrophic’ loss of control, says board member
- Anthropic researcher quits with a warning: Self-improving AI could "kill us all"
- Anthropic researcher believes more than 10% chance AI 'could kill all humans'
Across the whole window
The two tables below describe every item in the last 365 days under existential-risk. They are deliberately not narrowed by the reading filter (currently strongly adverse), because a mean taken only over items already selected for their score would just restate the filter.
By topic
| Topic | Branch | Items | Mean | Position |
|---|---|---|---|---|
| Existential and catastrophic risk | risk | 32 | -50 | |
| Concentration of power | risk | 23 | -28 | |
| Governance and regulation | structural | 22 | -20 | |
| Loss of control and alignment | risk | 22 | -63 | |
| AGI and timelines | structural | 21 | -37 | |
| Cyber capability | risk | 10 | -45 | |
| Evaluation and measurement | structural | 10 | -22 | |
| Human agency and epistemics | risk | 9 | -17 | |
| Scaling and architecture | structural | 9 | -8 | |
| Synthetic media and manipulation | risk | 6 | -63 | |
| Medicine and health | benefit | 5 | +0 | |
| Biological and chemical uplift | risk | 4 | -20 | |
| Scientific acceleration | benefit | 4 | +17 | |
| Education and access to expertise | benefit | 3 | -24 | |
| Abundance and material wellbeing | benefit | 2 | -8 | |
| Climate and energy | structural | 2 | -6 | |
| Compute and infrastructure | structural | 2 | -13 | |
| Employment and displacement | risk | 2 | -4 | |
| Productivity and growth | benefit | 1 | -26 |
By source
A source's mean says how the items it published were read. It is not a rating of the source.
| Source | Kind | Tier | Items | Mean |
|---|---|---|---|---|
| arXiv cs.CY (Computers and Society) | academic | 1 | 14 | -20 |
| The Guardian: Artificial Intelligence | news | 1 | 9 | -55 |
| Financial Times: Artificial Intelligence | news | 1 | 8 | -53 |
| The New York Times: Technology | news | 1 | 7 | -31 |
| BBC News: Technology | news | 1 | 6 | -36 |
| Ars Technica: AI | news | 1 | 5 | -73 |
| OpenAI | lab | 1 | 3 | -17 |
| IEEE Spectrum: AI | news | 1 | 3 | +2 |
| arXiv cs.AI | academic | 1 | 3 | +6 |
| MIT Technology Review: AI | news | 1 | 2 | -50 |
| The Economist: Science and Technology | news | 1 | 1 | -16 |
| Google DeepMind | lab | 1 | 1 | +7 |
| CSET, Georgetown | government | 1 | 1 | -22 |
| arXiv cs.LG (Machine Learning) | academic | 1 | 1 | +4 |